View Single Post
  #1977 (permalink)  
Old 03-22-2011, 12:40 AM
[ACL]'s Avatar
[ACL]
VIP Member
Offline
Location: NY
 
Join Date: Feb 2010
Posts: 1,534
Reputation: 6350
[ACL] is a trusted member of the community[ACL] is a trusted member of the community[ACL] is a trusted member of the community[ACL] is a trusted member of the community[ACL] is a trusted member of the community[ACL] is a trusted member of the community[ACL] is a trusted member of the community[ACL] is a trusted member of the community[ACL] is a trusted member of the community[ACL] is a trusted member of the community[ACL] is a trusted member of the community
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Re: NAND Boot Testing - 03-11: FRX05 on NAND (data working again!)

Quote:
Originally Posted by natemcnutty View Post
Sounds great. Just curious on this, why don't we just cp the modules into system instead of symlinking from data? I think the issues we were having before probably had to do with mounting system as ro before android started.
i think originally it was because of size. Modules are big so we put them in data. Also because if we copy them and upgrade a kernel, we will overwrite the original ones. and cant revert. And if we keep coping with multiple kernel support then we balloon up system.

Or we can go the cyanogen route and mount em compressed as files.

Code:
# Mount compressed filesystems
mount squashfs loop@/system/lib/modules/modules.sqf /syste/lib/modules ro

mount squashfs loop@/system/xbin/xbin.sqf /system/xbin ro
However they have sqsh4 support on their kernel and we only have 3. So making them is a bitch. But can be done. Good news with this is that deploying modules is a breeze since its just one file we need to read during boot. The above line gets added to the rc file and voila, modules are solid. He even mounts his busybox that way. lol.
__________________
Reply With Quote